Note: This bug is displayed in read-only format because
the product is no longer active in Red Hat Bugzilla.
RHEL Engineering is moving the tracking of its product development work on RHEL 6 through RHEL 9 to Red Hat Jira (issues.redhat.com). If you're a Red Hat customer, please continue to file support cases via the Red Hat customer portal. If you're not, please head to the "RHEL project" in Red Hat Jira and file new tickets here. Individual Bugzilla bugs in the statuses "NEW", "ASSIGNED", and "POST" are being migrated throughout September 2023. Bugs of Red Hat partners with an assigned Engineering Partner Manager (EPM) are migrated in late September as per pre-agreed dates. Bugs against components "kernel", "kernel-rt", and "kpatch" are only migrated if still in "NEW" or "ASSIGNED". If you cannot log in to RH Jira, please consult article #7032570. That failing, please send an e-mail to the RH Jira admins at rh-issues@redhat.com to troubleshoot your issue as a user management inquiry. The email creates a ServiceNow ticket with Red Hat. Individual Bugzilla bugs that are migrated will be moved to status "CLOSED", resolution "MIGRATED", and set with "MigratedToJIRA" in "Keywords". The link to the successor Jira issue will be found under "Links", have a little "two-footprint" icon next to it, and direct you to the "RHEL project" in Red Hat Jira (issue links are of type "https://issues.redhat.com/browse/RHEL-XXXX", where "X" is a digit). This same link will be available in a blue banner at the top of the page informing you that that bug has been migrated.
Red Hat,
Please consider update nvme-cli to v1.8. We have seen a number of issues related to FC and ANA with v1.7.
Note from James Smart:
Given the number of pulls I've seen recently of 1.7 of nvme cli, and
given 1.7 has the regression for fc that stops all connections - I
pinged them to create a new kit with the fc fix. They released a 1.8
kit today with the fc fix and other critical bug fixes.
So if you talk to distros, etc - please highlight 1.8 nvme cli support
for fc and ana.
----------------
thx
Laurie
On 4/9/2019 2:27 PM, James Smart wrote:
There have been several regressions caused by upstream changes to
nvme-cli. They were done by the folks adding support for "persistent
discovery controllers" which seemed to be paired with adding
nvme-over-tcp support to the cli. Many of these changes were done a
while ago - last Oct or Nov.
I know of the following issues:
1) a regression where the nvme cli will no longer connect to anything
on FC. The regression was added in rev 1.7 of the cli. A patch was
sent a while ago, but a "release" of the cli with the fix is only
available as of yesterday in rev 1.8 of the cli. If any user is
using rev 1.7 vanilla, they must apply the following patch to be
successful:
https://github.com/linux-nvme/nvme-cli/commit/324fce3d6adacdd0aad37e3b1962ab59f601050e
2) a regression where nvme cli will stop it's connect-all loop if it
encounters a discovery log entry that cannot be connected to. The
regression as added in rev 1.7 of the cli and continues to exist in
1.8 of the cli. A fix was posted upstream today, but is not yet in
the nvme-cli repository. Description: Anything that was in the log
up to that point would be connected. Anything after that would not
even be attempted to be connected to. It appears this is a common
thing on NetApp arrays. Their discovery log contains all target FC
ports for the array and it's not uncommon for an initiator to not be
zoned with all ports on the array - so its not uncommon to encounter
failures connecting to a log entry (which is the following dmesg string:
Failed to write to /dev/nvme-fabrics: No such file or directory"
and kernel: nvme_fc: nvme_fc_create_ctrl:
nn-0x200000a0986a8b5f:pn-0x200400a0986a8b5f -
nn-0x200000109b3c081f:pn-0x100000109b3c081f combination not found
Any entries in the log would not be connected to - so nvme controllers
won't be seen for them.
A fix for this error is to build the cli with this patch:
http://lists.infradead.org/pipermail/linux-nvme/2019-April/023319.html
3) a regression where the cli will not connect to longer subsystem
lists. Prior to nvme cli 1.7, the cli read the discovery log as a single
large read. Unfortunately, reads can be constrained in size by the
transport hardware (ex: lpfc pre 12.0.0.7 had a 256k limit) or by the
MDTS of the target (ex: initial NetApp array capped ios at 64k). Which
means - if the discovery log is sufficiently large, it can't be fully
read by nvme cli. nvme cli 1.7 was updated to read the log in multiple
smaller (4k) reads. Unfortunately, the nvmet core layer didn't
recognize that "offsets" were used for multiple reads of the log, so an
attempt to use 1.7 on a larger discovery log will result in erroneous
information (Note: Justin also says the xgig analyzer doesn't understand
these multiple-small reads either). The patches for nvmet core were
submitted upstream this week but it may be a while until a kernel with
the fixes exists in the distros. Note: NetApp successfully allows
these multiple-small reads.
the upstream patch for this error in nvmet core is: (v3)
http://lists.infradead.org/pipermail/linux-nvme/2019-April/023317.html
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.
https://access.redhat.com/errata/RHBA-2019:3545